For unmanned aerial vehicle (UAV) systems, efficient image transmission is critical for real-time applications, such as disaster recovery/relief, surveillance, and environmental monitoring, particularly in scenarios with limited bandwidth. To this end, this paper proposes a data-efficient image transmission pipeline to consider UAV systems’ transmission speed and storage capacity jointly. The pipeline compresses each image captured by the UAV using color quantization into a 16-color palette. This approach reduces image sizes considerably, facilitating quick image transmission over limited communication links. At the ground station, reconstruction of the image quality is pursued using deep learning-based CCDNet and super-resolution models to improve its perceptual fidelity. In contrast to the prior work, our approach is the first to uniquely introduce both quantization-aware restoration (CCDNet) and perceptual super-resolution (ESRGAN), allowing for the restoration of color fidelity and fine texture details for significantly compressed UAV images. We have demonstrated the performance of the proposed method on a dataset of 400 images that were taken by a UAV from different landscapes, and achieved significant data reduction in image size (~68%), resulting in ~3x faster transmission speed and significantly increased onboard storage capacity. In addition, qualities of restored images were presented with satisfactory SSIM (0.9419) and PSNR (34.90 dB). These results demonstrate the potential of the proposed pipeline in efficiently improving live image transmission and onboard storage capacity in UAV applications.